Energy intake and dietary macronutrient composition is the most difficult part of the energy balance equation to measure. In a natural history study of factors which predict food intake and using an inpatient computerized vending machine system, food intake has been measured over 300 individuals. In individuals undergoing the study more than once the intra-class correlation coefficient is very high (r=0.9) indicating that these studies although performed in an inpatient setting are very reproducible. This study has demonstrated that higher respiratory quotient, the ratio of carbohydrate oxidation to lipid oxidation which predicts weight gain, also predicts energy intake. The component of respiratory quotient which best predicted food intake was higher carbohydrate oxidation indicating a role for carbohydrate (likely glycogen) balance in the regulation of short term food intake. However, we did not find that glycogen depletion increased food intake in a follow-up study. We have also demonstrated that one of the most important determinants of food intake is fat free mass. There is a strong positive association between fat free mass (adjusted for height) and food intake, which persists even when adjusted for fat mass. Fat free mass is the major determinant of energy expenditure, and 24-hour energy expenditure is also strongly associated with energy intake. Furthermore, even after adjustment for fat free mass, the residual of energy expenditure was still associated with energy intake. In mediation analysis, energy expenditure accounts for over 80% of the explained variance in energy intake indicating that energy expenditure rather than fat free mass drives energy intake. Thus, increases in metabolic rate (as a weight loss intervention) may have the paradoxical effect of stimulating excess caloric intake. In fact, in further analyses, we have found that energy intake adjusted for energy expenditure predicts weight gain. We have begun investigating this energy sensing link further in a protocol which investigates if increased energy expenditure (as induced by cold) will lead to increases in food intake. In this approved protocol we are investigating also the timing of the increase in energy expenditure, whether this energy sensing occurs immediately or following the cold induced increase in EE. This protocol is currently actively recruiting participants. Using this ad libitum vending machine model, we have found that calories from soda intake predicted future weight gain indicating a role for sweetened beverages as contributes to the current obesity epidemic. We have also recently found that individuals who select a higher percentage of foods in high simple sugar/high fat content category gain more weight. In an attempt to understand hormonal underpinnings for macronutrient selection, we investigated the role of plasma GLP-1 and urinary dopamine concentrations on food selection. We found that increased GLP-1 concentrations were associated with lower carbohydrate and lower intake of high simple sugar/high fat foods indicating a role for GLP-1 in rewards based food selection. Urinary dopamine was associated with greater drive to eat and overall increased energy intake. In an effort to understand behavioral and psychosocial factors, we have also investigated whether socioeconomic status and food insecurity are associated with energy intake. We found that those with higher food insecurity scores, eat more during the ad libitum period indicating an important psychosocial component is at work in determining calorie intake. We have completed continuing our initial study investigating stable isotopes as biomarkers for dietary patterns. Individuals (n=32) consumed diets over 3 months in a multi-factorial design which varied by meat, soda and fish content. Plasma (collected every 2 weeks) and hair, were analyzed for changes in the stable isotopes C13 and N15. In both plasma and hair, carbon isotope enrichment was associated with meat and soda intake, but more strongly with meat intake, while nitrogen isotope enrichment was associated with fish intake. Of interest, pre-intervention isotope ratios were associated with post-intervention ratios indicating that even after 12 weeks of a stable diet, that turnover was incomplete. We will also investigate how these dietary patterns affect behavioral and performance testing, glucose tolerance and energy expenditure. We are currently planning to perform further measurements on isotope ratios of specific amino acids (e.g. alanine) which may be more specific markers of soda intake. We are also planning to perform analyses of fatty acid profiles on plasma and adipose tissue and metabolomics in plasma to further develop dietary signatures. We also have collected stool samples and will investigate whether gut microbiome changes may be associated with changes in circulating biomarkers.